Regulating Crime Hot Spots:
A Bio-Social Perspective on Policing
Lawrence W. Sherman
University of Cambridge
2010
Regulating Primates
Jessica Flack, D.C. Pigtail Macaque
Krakauer & F.B.M. de
Waal (2005)
“Robustness
Mechanisms in
Primate Societies: A
Perturbation Study”
Proceedings of the
Royal Society London
B 272: 1091-1099.
Overview of Primate Regulation
Frans de Waal
Two Lateral Questions
1. Is regulation instinctive in humans?
2. Does regulation prevent violations, or
merely case displacement of them
from one venue to another?
Thesis
A bio-social framework for the imposition of, and
compliance with, regulation will help improve regulatory
sciences in general, and in particular will…
1. test the displacement hypothesis in a broad range of
regulated behaviours
and perhaps…
2. help to falsify displacement as an “instinctive” reason to
prefer retribution over regulation.
1. Is Regulation--Like Revenge--
Instinctive?
• Diamond: Guns, Germs and Revenge
• Neuroscience: Vengeance is Sweet
• Kant: Retribution is “Natural” Justice
• Elephant “payback” attacks on human villages
• Why not regulation (or even forgiveness)—as
well?
So What?
Policy Debates
We are “Hard-wired” to do X, Y or Z
• Implies functionality—if not inevitability
• “Common Sense”
• “What Works”
• Opposition is futile, idealistic
• It is “human nature” to be …….
• “Survival of the fittest”—laissez faire policy
• Revenge is inevitable—we must feed it
Restorative Justice Reduces
Desire for Revenge
Sherman, Strang et al 2005
Biology vs. Bio-Social Theory
• “Instinct” is pure biology—perhaps for lice
• Social animals are also shaped by social
environment
• Two may fuse, as in evolution of Bonobos
• Scarce resources for Chimps
• Not for Bonobos
• Which species became more violent?
Bio-Diversity Variability
• 20% of humans have no benefit from
cardiovascular exercise
• Some humans allergic to penicillin, peanuts
• Others sensitive to caffeine, alcohol
• SEX:
--5 to 10% of humans (?) have same-sex
orientation
--Even more humans = no-sex orientation
--Some species reproduce without mates by
parthenogenesis
But Note the Policy Difference
That Biology Makes
1. Homosexuality de-criminalized when
bio-evidence supported “instinctive” view
2. Gender equality promoted when women
tested in field as police, executives, etc.
3. Murderers are spared from execution
when brain structures shown to be
defective
Diversity and Consistency
• Two themes of biology
• Both providing assurance of what is “natural”
• But both also adding to explanatory power
• Bio-social view may improve theory
• Provide consilient explanations: when
predictions from one class of facts are confirmed
by predictions to another class of facts
• Consilience is the test of a data-integrated
theory (vs. a mere hypothesis)
Virtues of Consistency
Consistency of variability, that is:
• Variability within individuals
• Within groups
• From peace to violence
• From revenge to reconciliation
• From low to high to low crime, violations
• With and without social & resource factors
Consistency Across Hominoids
• Humans
• Bonobos
• Chimpanzees
• Orangutangs
• Gorillas
All evolved to live in groups, cooperatively
And all have inter-personal conflicts regulated by the group
All may provide clues to bio-basis of human behavior
Primatology: Regulated
Competition & Cooperation
• Life in small groups
• Cooperation for food
• Altruism across all 5
species
• Hierarchies of
dominance
• Used to regulate
competition
• In interests of
cooperation
Variable Patterns of Violence,
• No good estimates of cross-species
homicide rates
• Intra-group lower than inter-group
• Chimps more inter-group than bonobos
• Bonobos: female dominance, execution
• Chimps: female restorative justice
• Chimps: alpha male policing
Chimpanzee Regulation
1. Pre-Violence
• Anger displays, noise, face-off
• Alpha males: “Eh, Wot’s All This Then”
• Face-off disbands
2. During Violence
• Alpha males use force, knock fighters off
3. Post-Violence
• Females pester males to groom each other
The Chimp Patrol Experiment
• Yerkes Observatory, Atlanta
• Alpha males regulating conflict
• Experiment: removal from group, 8 hours
• Visible through clear glass wall
• Chimps can see “police” are incapacitated
• Result: all hell breaks loose
• Alpha males return restoration of order
Causal Inference?
1. Chimp “police” general deterrence
2. Chimp police individual deterrence
3. Chimp police forcible restraint of harm
Predicting Across Groups
• Police strikes predicted chimp result
• Whenever police removed from cities
--Montreal
--Helsinki
--Others
• Can chimps predict human results?
• What about displacement?
Two Theories of Crime
• Individual Causation:
Like a meteor, criminal will keep flying until
energy is all used up, committing a fixed number
of crimes before flaming out—if not here, then
somewhere else
(serial murderers, Ponzi scheme artists)
• Social Interactions
Like a gas stove, offenders will keep re-
lighting, but only if a) gas is on, and b( someone
lights a match
Routine Activities Theory
• Cohen and Felson ASR 1979
• Criminal events require 3 conditions
1. Absence of capable guardians
2. Suitable target
3. Motivated offender (absence of handlers)
Hot Spots Patrol Experiment
(Sherman & Weisburd, 1995)
• Concentration of violence, human space
• 3% of addresses = 50% of incidents
• 100 hot spots:
55 = 15% of time had uniformed police presence
55 = 7% of time had uniformed police presence
• 2/3 reduction in total incidents
• 50% reduction in robbery
• 50% reduction in anti-social behaviour
Big Question: Displacement
• Of EVENTS to adjacent locations
• Of OFFENDERS to any other location
1. Hot Spots of Crime and Disorder
What is a Hot Spot?
• Concept
• Varieties of Definitions
Broadest Definition
A hot spot is a geographic space in which
there has been a concentration of crime
(however measured) per square foot
relative to other space in the larger
jurisdiction.
What Makes a Hot Spot Hot?
• Density of people
• Density of risk factors—alcohol, weapons
• Density of targets—cars, drunks, cash
• Structure of routine activities--unguarded
Philadelphia Hot Spots
Peaks and Valleys of Crime
(Distribution of Violent Offenses in Tokyo)
Homicides by Address
NW Washington DC
Visualization of Spatial Data
(Distribution of Violent Offenses in Central 23 Wards of Tokyo)
15-Year Trajectory of Blocks
Crime Drop Concentrated
• Most of the drop in crime
• City-wide
• Concentrated in the 15% of street
segments
• With most crime
• Suggesting non-displacement
COMPSTAT implications: Time
• Why look at one month?
• Why not one year?
• Ten years?
• What can you do differently with
trajectories?
• COMPSTAT—how does each district use?
• COP-stat—how many police in each
district?
What is Effect of Different Policing
Strategies in Hot Spots?
• Marked car patrol
• Drug Raids
• Field Interrogations
3. Uniformed Patrol
• Minneapolis Preventive Hot Spots Patrol
Experiment:
• A Randomized Controlled Trial
Minneapolis Hot Spots
• 1988-89
• 110 Intersections, randomly assigned
• 55 got 3X patrol
• 55 did not
• Difference held for 8 months
• 2/3 less increase in calls for crime
• 50% less Anti-social behavior--observed
Outcome Measures
• Calls for Service (CHEAP)
--fights, thefts, disputes, noise, robbery,
assaults
• Observed disorder (DEAR)
--drunks down, prostitution, drug sales,
aggressive panhandling
Control Group, Patrol Group
• Crime Went Up in Both
• Much more in Controls
• Difference: Caused by extra patrol
Results
• 2/3 fewer calls for service relative to
controls
• 50% less robbery
• 50% less observed disorder
Displacement unlikely
Experimental Hot Spots Improved
Relative to Control Hot Spots
30%
25%
20%
Change 15%
In
Crime 10% Control
Exp.
5%
0%
-5%
-10%
Total Calls Soft Crime Hard Crime
Koper Curve
• Dynamics of RESIDUAL deterrence
• Optimal Time for Patrol Presence
• 10-15 minutes
• Defined by time to next crime observed
• After 15 minutes diminished returns
• Maximum effect: ROTATION
Optimal Patrol Time Per Hot Spot:
10-15 Minutes (Minneapolis)
• Minutes police
Present
• Minutes to
first crime after
Police leave the
Hot Spot
Effects on Crime and Disorder Maximized by 14-
15 Minute Stops: Koper (1995) Curve
1.6
1.4
1.2
Effect
On 1
Crime
And0.8
Disorder
0.6
0.4
0.2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Length of Stop
Source: Koper (1995)
Florida Test
• Unpublished
• PERF
• Koper, 2009 (Cambridge Conference)
Florida: Preliminary Results
(Reductions in serious non-domestic violence at 140 days)
30%
25%
20%
15%
10%
5%
0%
Saturation Problem Solving Control
Based on before and after tracking by JSO
Greater Manchester Police:
Tactical Experiments and
Strategic Testing (TEST)
Extending the US Hot Spots
Patrol Experiments 1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Research Design: GMP
• Identify 200 Hot Spots small enough so that one police
car can be seen from anywhere in the hot spot
• Make each hot spot far enough from other hot spots so
police cars in one cannot be seen in other hot spots
• Add a “displacement cushion” around each hot spot to
see if crime moves around the corner
• Assign 100 hot spots to 2 hours daily of extra PC
patrol
• Assign 100 hot spots to standard PC coverage (no
directed patrols
• Compare before-after crime trends
Randomized Controlled Trial RCT:
COMPARISON or NET difference
140%
120% 121%
Court
101% Offender
100% 100%
s (n=59)
80%
71% 71%
DC
60% Offender
s (n =
49%
62)
40%
28%
20%
Yr (-2) Yr (-1) Yr (+1) Yr (+2)
Treatment Measures
• Minutes of PC presence in each hot spot
as measured by ARLS
(automatic radio locator system)
• Arrests made in each hot spot
• Qualitative observations in PC, PCSO,
and control hot spots
• Other police forces welcome
Outcome Measures
• Incidents of public crime, anti-social
behaviour and disorder as reported by
telephone to police
• Crimes occurring in public for which
written reports are taken by police
• Quality of police-community interactions
in three treatment groups
• Diploma, MSt
• Applied Criminology &
Police Management
• Senior Command
Course
• Chief Inspector to CC
• Crime analysis:
Analysts and Sworn
Alfred North Whitehead,
1861-1947
“Universities
[help]
create the
future”
Yeats: “Education is not the
Filling of a Bucket, but the lighting of a fire.”
“Crackdowns”
• Sudden Increases in Police Patrol or
Arrests
• Targeted on Areas or Offense Types
• Generally of short duration
• Common for drugs, drunk driving
• Was Petra Todd’s study really a
crackdown?
Crackdown Concepts
• Initial Deterrence:
drop in crime
• Deterrence Decay:
return of crime towards previous levels
while crackdowns continue
• Residual deterrence:
crime stays lower even after crackdown
ends
Crackdown—Backoff
Residual, Decay
4. Drug Raids
• Drug Dealing blocks—Kansas City
• Elevated crime, disorder
• Crime on the block
• Driven by drug house; complaint number
RCT of Drug Raids
• Hotline complaint
• Undercover officer
• Drug Purchase
• Search Warrant
• Random assignment
• 50% raided
• 50% wait 30 days
• A “wait-list” trial (Sherman & Rogan, 1995)
5. Evidence on Stop, Search
• Kansas City Gun (1)
• Indianapolis (2)
• Pittsburgh (2)
• Colombia (2)
Stop, Search and Gun Violence:
The Carrying Hypothesis
Lawrence W. Sherman
Summary
• Weapon violence depends on weapon
density
• Density control may be best prevention
• But given density, carrying may be best
point of intervention by police
• 8 out of 8 tests of hypothesis
• Suggests policing can reduce gun violence
• Likely causal mechanism = public carrying
Kansas City Example
• Top Drive-by shootings Police Beat, 1991
• Homicide rate 177 per 100,000 residents
• US homicide rate about 9 per 100,000
• Difference in Rate 20 to 1
• KC rate 31 per 100,000
• Within-city difference = 6 times higher
• Add time of day difference of 5:1
Kansas City Gun Experiment
• 4,500 hours of extra patrol paid for
• 1,218 hours delivered
• 1,090 traffic citations, 532 ped. Checks
• 616 arrests
• 29 guns seized
45% search incident to arrest
34% frisk (Terry v. Ohio)
21% plain view
Systematic Review
(Anthony Braga, Harvard)
• IN: All tests of police activity to detect guns
being carried illegally in public places
for effects on gun injury, homicide
• OUT: other deterrent efforts against gun crime
--Bartley Fox law (Mass)
--Boston Gun Project
• Koper& Mayo-Wilson JEC—same conclusions
Eight Independent Tests:
• Sherman & Rogan 1995 (1)
• McGarrell et al 2001 (2)
• Villiveces et al 2000 (2)
• Sherman 2000 (1)
• Cohen and Ludwig 2002 (2)
8 of 8= Outcome was
Homicide or Gun Wounds
• Kansas City, Mo. Homicide
• Indianapolis Homicide
• Indianapolis Homicide
• Bogota Homicide
• Cali Homicide
• Pittsburgh Wounds
• Pittsburgh Wounds
• USA Homicide
Percentage Differences
• Kansas City, Mo. 73% reduction
• Indianapolis 86% reduction
• Indianapolis 100% reduction
• Bogota 8.5%
• Cali 17%
• Pittsburgh 73%
• Pittsburgh 73%
• USA 22%
Gun Seizure Difference
• Kansas City, Mo. 65%
• Indianapolis 8%
• Indianapolis 49%
• Bogota ?
• Cali 500%
• Pittsburgh no
• Pittsburgh no
• USA 150%
Guns Found May not Matter
• Consistent Effect from LOOKING for guns
• Not actually finding them
• What matters may be threat message
• Questions about communication
• Pittsburgh—day-to-day differences
• Gun Carriers =
Sophisticated consumers of police work?
Proceed with Caution
• Still possible chance effects
• Regression to mean
• No randomized controlled experiment
• No clear theory of what works
But Do Proceed
• Absent any better evidence
• Combined with general evidence on patrol
• Best use of Police resources to reduce
gun crime
• No comparable evidence on other tactics
• Gun dealer interventions?
• Black market suppliers?
• Plan, implement and test
Knife Crime Analysis
• Not how many knives seized
• Where
• When
• Proportion seized proactively
• Proportion seized in hot spots
• Potential for a safer city—moderate at best
• Longer term solutions still needed
6. Evidence on Displacement
• CRIMINOLOGY 2006
“DOES CRIME JUST MOVE AROUND THE
CORNER? A CONTROLLED STUDY OF
SPATIAL DISPLACEMENT AND
DIFFUSION OF CRIME CONTROL
BENEFITS”
DAVID WEISBURD, et al
Kansas City Gun Experiment
• Small increase in catchment area
• Not statistically significant
• Much larger catchment area than target
7. Unanswered Questions
• Dose-Response curve
• People v. Places displacement
• People v. places legitimacy
Does-Response Curve
• How many more crimes prevented
(what type, cost)
• Each additional patrol hour
• In Hot Spots
• Based on experiments, not correlations
People v. Place Displacement
• Moving people to other places
• Do they take crime with them?
• Seattle? (No)
• Out of Seattle?
People v. Place Legitimacy
• Cure crime, disorder
• Places and overall
• Effects of strategies on legitimacy
• People pushed away from places
Regulating Crime Hot
Spots:Policing
Lawrence W. Sherman
University of Cambridge
2009